Military Review

Robot games: how the Pentagon defeated the Russian army in 2030



Blue vs red

The hypothetical war between the Russian and American companies, according to the calculations of the analytical company RAND, took place in the early 2030s near the Latvian town of Gulbene. The Russian army was represented by the "red" unit, consisting of three motorized rifle platoons on the BMP-2M and one tank platoon on the T-90M. The unit was covered from the air by one Tunguska, and an artillery observer based on the BMP-2 with drone "Aileron". For what reason, American analysts chose just such an arsenal, which is not the most modern even for the present time, is unknown. It is likely that there was no suitable software, or they simply decided to pre-program the result of the battle.

Medium-class combat drone RCV-M. Potential participant in the conflict with Russia in 2030. Source:

It is worth mentioning separately that each combat vehicle was virtually controlled by a separate gamer operator. Only the actions of fully autonomous machines were partly due.

The American side or "blues" were represented by much more advanced armored vehicles. Of the “classics”, only the Abrams M1A2C remained, equipped with all the active and passive protection systems possible for its time. By the beginning of the 30s, according to RAND, the Pentagon will be armed with an optionally controlled BMP Optionally Manned Fighting Vehicle (OMFV). From weapons - 50-mm cannon and anti-tank missiles. The level of booking OMFV is rated at an average level. The BMP in the shock company was presented in two guises - six vehicles were with a crew and infantry, and two were part of a platoon robots. Six Robotic Combat Vehicle—Medium-Weight (RCV-M) were also present in the robotic platoon. In these vehicles, the armor can withstand the fire of a heavy machine gun, a 30-mm cannon and two ATGMs from weapons. But with the robots, the "blue" is not all. The light class vehicles were six wheeled/tracked Robotic Combat Vehicle—Light-Weight (RCV-L), each with a heavy machine gun and an anti-tank Javelin. Air reconnaissance was carried out by small-sized Drones Coyote with thermal and IR cameras. Such a mixed mechanized-robotic company was supposed to attack the Russian unit, drive it out and move east.

Lightweight robots RCV-L. Source:

The game had two scenarios (gameplay). In the first stories against the "red" were exclusively remote-controlled vehicles in conjunction with traditional technology. In the second case, technology with the rudiments of artificial intelligence and machine learning entered the battle. Such ground drones can independently determine the route and capture targets. Autonomous robots, in accordance with the scenario, had to obey the commands of the company commander and even open fire on their own, if they saw fit. The Pentagon has no plans to adopt such fully autonomous vehicles yet, but in the foreseeable future they may become operational. Interestingly, RAND analysts do not foresee the appearance of heavy-class autonomous vehicles in the American army by the beginning of the 30s. That is why the good old Abrams played the role of the main striking force in the game. There are several reasons for this disbelief in the technological progress of the United States defense complex. Firstly, robots on the battlefield are viewed as a bargaining chip, which is not particularly a pity. And the loss on the battlefield of a heavily armored robot worth several million will be noticeable for the military budget and the taxpayer's pocket. Secondly, the error of a robot with a large cannon is much more sensitive than a random burst of 30mm or 50mm guns from medium and light ground drones.

What happened in the end?

In fact, RAND started the game to compare the capabilities of remotely controlled drones and fully autonomous vehicles. With the Russian troops, despite the obvious backwardness for the 30s, everything was not so simple. The players equipped the infantrymen with knapsack jammers for radio control signals, which seriously complicated the work of unmanned remotely controlled vehicles. In this regard, autonomous robots were more stable, although they also had to receive separate commands from the company commander through the radio channels of vulnerable communication. In general, the conclusion of the American side was simple - something had to be done with Russian electronic warfare systems. They cannot disrupt the offensive of the company, but they can seriously complicate the control of the drones.

In any war, robots must be at the forefront of the offensive to conduct reconnaissance, save lives, and strike at the most defended areas of the front. In the game RAND made no exceptions, and the RCV light wheeled robots with medium crawler tracks were the first to attack. In fully autonomous mode, they turned out to be great baits! Defending Russians Tanks and infantry fighting vehicles opened fire on advancing drones, thereby unmasking themselves. And on the discovered firing points of the "red" American tanks and the attracted artillery were beating unanimously. From here, RAND concluded that fully autonomous drones were used primarily for opening the firing points of the defending enemy. Light cars are especially good in this role - they are like weapon offensives are not very good (machine guns are almost useless in classic combat), and it's not a pity. Analysts especially emphasize that such equipment cannot be sent to the vanguard of an offensive without aerial reconnaissance and artillery support. Drones, due to the weakness of surveillance equipment, will not really detect anyone and will quietly and peacefully die.

As a result, researchers at RAND and Forbes, inspired by the results of a military computer game, recommended that the Pentagon, on the readiness of autonomous robotic systems, integrate them into the battle formations of the army. Russian electronic warfare systems in the near future are quite capable of rendering useless all remotely controlled combat vehicles. The advice of the respected RAND team in the United States is usually heeded in the defense department. Analysts even named the date for the coming of a happy autonomous robotic future - the beginning of 2031. But not everyone agrees with this.

RAND headquarters in Santa Monica. Source:

Resource, following the analysis of the RAND games, released a large analytical material in which it predicted the long-term involvement of land drones in the fighting of the American army. The authors rightly point out the lack of serious combat experience in ground robots, even on remote control, not to mention fully autonomous systems. So far, the rudiments of artificial intelligence and machine learning can be found in flying reconnaissance and strike drones, but the specifics of their work are seriously different from ground vehicles. First of all, such wheeled or tracked "smart guys" should inspire confidence on the part of soldiers and officers. It is one thing when a robot is under human control, and quite another when it is “on its own”, and even with a gun. RAND's opponents cite a typical example of human perception of the first elevators in history. The inhabitants flatly refused to move between floors without an elevator on board. It took several decades of practice for us to feel confidently one-on-one with the elevators. With self-guided combat systems on land, the story will be similar.

Actually, there is still no clear understanding of what real autonomy is. Can we consider an autonomous machine that completely replaces human capabilities, or should a robot necessarily surpass human capabilities? Critics of the RAND concept generally suggest not to suffer with the rapid introduction of autonomous intelligence in ground combat drones until they are capable of full-fledged raids deep into enemy territory. Artificial intelligence must fight in close connection with winged drones, as well as constantly interacting with human-controlled combat drones. There is no need to wait for such technologies by 2031, so all attempts to partially turn on the autonomous strike "infantry" will only be a waste of time and money.

Subscribe to our Telegram channel, regularly additional information about the special operation in Ukraine, a large amount of information, videos, something that does not fall on the site:

Dear reader, to leave comments on the publication, you must sign in.
  1. nikvic46
    nikvic46 April 10 2021 06: 26
    How should you feel about such games? Someone seriously thinks that he will sit at the control panel and control the course of the battle. Recently, there were NATO exercises in Norway. They probably also worked out the course of the exercises. But after that, it turned out that the fighters did not have warm socks. And life brought these warriors to the ground.
    1. Cube123
      Cube123 April 10 2021 08: 21
      Quote: nikvic46
      How should you feel about such games?

      I take it seriously. Real teachings are needed to identify "little things" such as your example of socks. It is better to learn about this during the exercise phase than in a real combat situation. Therefore, after the end of the exercise, reports are always written and all situations that have arisen are carefully analyzed. So here, if you do not engage in such "games", then a lot of important points will be out of your attention - you simply will not know about them. And modern computer simulators allow simulating rather difficult situations. Yes, this is not a real war, but it allows you to identify real problems, determine the methods of training operators, and finally allows them to develop the necessary conditioned reflexes. This is much cheaper than real "games" in the field and saves lives and real expensive equipment.
    2. Installer
      Installer April 10 2021 09: 30
      Oorfene Deuce and his wooden soldiers. The first use of combat ground robots))
      1. 72jora72
        72jora72 April 11 2021 16: 04
        Oorfene Deuce and his wooden soldiers. The first use of combat ground robots))
        And the first robot with artificial intelligence (golem) was invented by the Jews lol
        in the Talmud they still wrote about them.
  2. Doccor18
    Doccor18 April 10 2021 06: 47
    Artificial intelligence must fight in close connection with winged drones, as well as constantly interacting with human-controlled combat drones. There is no need to wait for such technologies by 2031 ..

    "Down and Out trouble started"...
  3. SpbGenn
    SpbGenn April 10 2021 07: 49
    The Pindrs have emailed our WWII tactics - reconnaissance in force. The scheme included light tank robots instead of light tanks. What a depth of military thought !!
  4. PSih2097
    PSih2097 April 10 2021 07: 50
    For what reason the American analysts have chosen just such, not the most modern arsenal even for the present time, is unknown.

    What was in ARMA, then they chose ... wassat
  5. engineer74
    engineer74 April 10 2021 10: 47
    Again, spherical companies in a vacuum ... What if the Reds decide to use a remote mining system, for example? The picture will change upside down wink
    It seems that it suddenly dawned on the "partners" that the endless cut at the Air Force was starting to look very suspicious! wassat
    We decided to take on the ground forces with robots ... In short: "Saw, Shura, saw!" © good
    1. tank64rus
      tank64rus April 10 2021 15: 29
      I agree, in addition to remote mining, there is also electronic warfare, the setting of masking and blinding curtains, engineering barriers, air defense systems for UAVs and aviation. Up to the interception of control of the enemy robots. And how the electronic stuffing of robots will react to the use of tactical nuclear weapons. There are more questions than answers.
  6. The popuas
    The popuas April 10 2021 11: 01
    Tunguska covered from the air wassat did not read further ...
    1. sedoj
      sedoj April 10 2021 14: 37
      Looks like "Tunguska" in a hot air balloon. As well as the "train with an aviation bias".
  7. Knell wardenheart
    Knell wardenheart April 10 2021 11: 04
    Computer vision is far from perfect, computer logistics algorithms tied to computer vision are also not without drawbacks. If in BB2 and interbellum they experimented with teletanks, then as a rule they faced the problem of unsatisfactory mobility and armor. This problem remained + the above + extreme dependence on sensors. Where a person with a flexible and creative mind and experience will not see a problem in the circumstances that have arisen - the machine can run into a wall, where it will fall for a penny.
    I mean that the front line of the defense will simply be equipped with the expectation of immobilizing or detaining such automated complexes. They will use a variety of means of dazzling them and illuminating the sensors, finally, their weak armor will also be a problem - including for the destruction of such robots using UAVs.

    I'm not sure that at the beginning of 2030 the situation will change dramatically - ground-based robotic systems seem to be stuck for a long time in the niche "object protection" / "demining" / "ambush operations" / "point sabotage".
  8. Operator
    Operator April 10 2021 13: 20
    REND is clearly out of control - if in 2030 an AI appears capable of autonomously controlling ground robots, it will all the more be able to autonomously control aerial robotic drones, since there are no rough terrain, mines, traps, camouflaged firing points, etc. in the air. etc.

    As a result, ground robots fired in front of the advancing infantry, worth a million dollars each, will not reconnoiter anything in the enemy's ground line of defense for a simple reason - they will use penny autonomous aerial robots in conjunction with 120-mm mortars, firing homing mines from closed positions, or those the same ATGM, also firing from closed positions.

    Even in the event of a tie in a battle between robots, the attacker will suffer multimillion-dollar damage, and the defenders - multi-penny damage.

    In general - Abrams kaput bully
  9. ximkim
    ximkim April 10 2021 14: 17
    According to RAND's calculations, full-fledged robotic and autonomous systems are not expected in the RF Armed Forces.
    1. Vadim237
      Vadim237 April 10 2021 15: 19
      It is still unknown what will be created in our country by 2031, but work on photonics, a quantum computer and AI are now going on throughout our country.
  10. Thomas N.
    Thomas N. April 10 2021 14: 22
    What an archaic suspension this RCV-M has:

    1. Couchexpert
      Couchexpert April 10 2021 18: 13
      Hmm ... Old suspension ... Or maybe "this is not a bug, but a feature"?

      1. Thomas N.
        Thomas N. April 10 2021 18: 30
        The M24 Chaffee and UAS Companion, shown by you, have a very modern suspension.
    2. Saxahorse
      Saxahorse April 10 2021 20: 42
      You showed a completely different suspension. The first is a typical tractor, the second is active, torsion bar.
  11. BAI
    BAI April 10 2021 19: 26
    Light vehicles are especially good in this role - they are not particularly good as offensive weapons (machine guns are almost useless in classic combat), and it's not a pity.

    And who said that you need to shoot at them from tanks and open your defenses? It is quite possible to interrupt them with grenade launchers, such as RPG-7. And then there are the means of suppressing optics. And where will the blind robot go?
  12. andrew42
    andrew42 April 10 2021 20: 03
    Robots and drones are of course fashionable and interesting. And if everything is more or less clear with drones (including ground ones) under the control of operators, then dancing with tambourines around "autonomous combat robots" with AI causes a feeling of delirium of "Novo-Tukhachevskys" coupled with the excitement of "sawmen". The concept of using "autonomous combat robots" is in any case tied to the definition of "friend or foe" as a fundamental task for AI. This problem is solved only by exchanging information (request - confirmation / ignore) between the active subject and the requested object. at least this happens between living fighters, and AI a priori operates within the framework of human logic, because it is created not by supertectones from Alpha Centauri, but by people from planet Earth. But it is easiest to provide "ignore" through the channels of perception. And how will there be guaranteed protection from the actions of an "autonomous robot" against its fighters / equipment, if he damn it "autonomous"? - AI itself must identify the "side" according to one of the implemented criteria, or it will have to be forgotten about the real "autonomy". So far, the "autonomous robot" with AI is quite capable of functioning in the "urine all objects similar to ..." mode, that is, all. A sort of Karbaras from the fairy tale "One-two, woe is not a problem." Until they turn it off. And so that theirs are not around And if the "control" is pulled out or blocked, then AI will turn out against everyone. Conceptually, this idea of ​​"autonomous" combat AI is delusional. You can grumble over this endlessly, but breaking all efforts is easier by a couple of orders of magnitude.
    1. Thomas N.
      Thomas N. April 11 2021 08: 00
      Everything is correct, so that ground-based autonomous robots with algorithmic similarity of AI should only be sent to the rear of the enemy to create chaos there and divert forces to their search and destruction. And genuine artificial intelligence (with self-awareness) will not be in a hundred years, and when it is created it is still unknown for whom it will want to fight - for us or against everyone in general, like Skynet - it will turn out to be such a "humanist" and will carry out an operation to "compel the world "for all mankind at once. wassat
  13. nikolai.shupenin
    nikolai.shupenin April 10 2021 20: 18
    none of this will happen. Europe will burn in a nuclear fire.
    1. Thomas N.
      Thomas N. April 11 2021 07: 22
      Quote: nikolai.shupenin
      none of this will happen. Europe will burn in a nuclear fire.

      Europe is from the Atlantic Ocean to the Ural Mountains, so this also concerns us.
  14. bk0010
    bk0010 April 11 2021 15: 45
    In fully autonomous mode, they turned out to be great baits! Defending Russian tanks and infantry fighting vehicles opened fire on the advancing drones, thereby unmasking themselves.
    Return a "boot" to each platoon and the problem will be removed. Only the positions of the grenade launchers will be revealed (the grenade launchers can immediately knock down after a volley, and the pipe is not very sorry).